Ant Group’s Ling 3.0 Flash packs 124B parameters into a model built for speed, not size

Share

Ant Group, the fintech giant behind Alipay, launched the Ling-3.0-Flash model on July 23, 2026 through its inclusionAI research lab. This open-source mixture-of-experts model contains 124 billion total parameters but activates only around 5.1 billion during inference. It matches or outperforms Ant Group’s previous flagship model Ling-2.6-1T, which had one trillion parameters, on core reasoning and instruction-following benchmarks. Ling-3.0-Flash scores 38 on the Artificial Analysis Intelligence Index and features a native context window of 262,000 tokens. The model is available under the MIT license on Hugging Face, with free API access through OpenRouter and Kilo until August 3, 2026.

Source: Read the original article

Telemac
Telemachttp://cryptoinfo.ch
Passionné de nouvelles technologies, j’explore l’univers de la blockchain et des cryptomonnaies pour partager l’actualité et les innovations du secteur.

Lire la Suite

Articles